How We Check Listings
Every listing starts from a page the program itself publishes. That page belongs to a state lab, a county or city health department, an extension office, or a water utility. We open it, and we copy a price, a test list, a phone number, or an address only when that page states it. The research file keeps a verbatim quote beside the fact. If the page is silent, the field stays empty. We do not fill the gap.
Each card shows the day we opened the official page. We open it again within 90 days. A recheck date that has already passed fails the build, so a stale card does not stay up on its own. When the official page says the program is sold out, suspended, or not taking orders, the listing is marked paused. It still appears on the state page with a warning, and search leaves it out.
When the live page would not open, the card uses an archived copy of that page and names the date of the copy. If the archive record has only a year, the card says the year.
No program pays to be listed or to rank higher. A search sorts by distance, then by price. The directory is kits and lab tests. It does not list equipment you borrow or rent.
Why the official page is the source
Private well owners carry the testing job themselves. EPA says: "Private well owners are responsible for delivering safe drinking water to their households." The same page says the quality and safety of water from private domestic wells "are not regulated by the Federal Government under the Safe Drinking Water Act nor by most state governments and laws." (EPA)
USGS Circular 1332 sampled about 2,100 domestic wells in 48 states. About one in five of those wells, 23 percent, had at least one contaminant above a human-health benchmark. The circular also says individual homeowners are responsible for maintaining the well and for monitoring water quality. (USGS)
That is why a county health department or a state lab page is the listing. Any price on the card is the price on that official page, not a national price we made up. EPA estimates that more than 23 million households rely on private wells. "Around 15 percent of the U.S. population (over 43 million) rely on private wells as their source of drinking water." (EPA)
City and town water is a different record. EPA requires community water systems to deliver a Consumer Confidence Report, also called an annual drinking water quality report, to their customers. "Reports must be sent by your water supplier each year by July 1." (EPA) A utility test listed here is the offer on that utility's own page, not a rewrite of the report.
EPA certifies laboratories that analyze drinking water samples for compliance with the regulations. (EPA) When a state publishes its own list of those labs, this site files it as a lab list. It is not a test you order from us.
Health limits on the results page come from EPA's national primary drinking water regulations, and each limit links to that table. (EPA)
What the numbers count
The home page, this page, the CSV download, and the counts file use one classification. A row is one published water program. Radon test kits live on the sister site and are not in these numbers.
- 330 public tests
- A water test you can order now from a state lab, health department, or utility. Lab lists, help paying for a repair, and paused programs are not in this number.
- 43 free tests
- Public tests whose posted price is $0 and that are open. Paused, seasonal, and limited offers are not counted as free. A test only for irrigation, or only while supplies last, is not counted as free.
- 46 certified-lab lists
- An official list of certified labs. It is not a test you order from this site.
- 5 paused programs
- The program page said sold out, suspended, or not taking orders when we checked. The state page still shows it, with a warning. Search leaves it out.
- 1 help listing
- A grant or similar help paying for a fix. Not a water test.
- 39 city and utility tests
- Public tests run by a city or a water utility. Statewide and county programs are in the public test count, and they are not in this number.
- 382 all records
- Every published water program, including paused programs, lab lists, and help listings. Radon test kits are on the sister site and are not in this count.
The CSV download has one row per published record. Its columns are state, name, agency, price, tests, official_url, last_checked, id, kind, availability, and free. kind is test, lab-list, or help. free is yes only for a public test counted above as free. id is the stable program id. Counts as JSON repeats the classes above.
Each published program keeps at least one verbatim quote from its official page in the research file. The card prints the fact and the day we opened that page. If the page does not state a price, a phone number, or a test, that field stays empty.
What the county tables count
A county page puts its water facts in one table. Each row is a number the source published for that county, with a bar drawn from it, and for most rows the same measure for the state beside it. The state figure adds up every county in the state where the source published that number, so a system listed in two counties counts in each, and the state column is a share, not a count. The definitions live on this page. A blank cell is left off the table. A published zero is shown. What the numbers count above is the program list, a different count. The household water, community water, PFAS, and system charts were checked October 11, 2026.
Household water
The household water row is USGS domestic self-supplied population for 2015, in people, out of the county population USGS used that year. USGS published the figure in thousands. The page rounds it and says "about." USGS did not publish a well count, and it did not split wells from other self-supplied sources. (USGS)
Community water
Where CDC published 2022 figures and they are the county's most notable finding, the row shows people on community water in the nitrate or arsenic bin CDC reported, and their share of everyone CDC counted on community water there. (CDC)
PFAS detections
A PFAS detection in the UCMR 5 row is a result at or above the method reporting limit. EPA collected those samples from 2023 through 2025. A county with no row in that older summary has no PFAS line in the table. A missing line is not a finding that the water has no PFAS. EPA's April 10, 2024 rule set a 4.0 ppt maximum contaminant level for PFOA and a separate 4.0 ppt level for PFOS. That is not a limit for every PFAS chemical. (EPA)
Violations
A health-based flag counts EPA ECHO systems with HEALTH_FLAG yes. The flag means a maximum contaminant level, a maximum residual disinfectant level, or a treatment technique. ECHO's help text does not give that flag a 3-year or 5-year window. The 3-year count is QTRS_WITH_VIO, and it counts every violation type. The extract date is July 9, 2026. (EPA ECHO)
The health-based violations row counts violations from October 17, 2021 through October 17, 2026. The pull date is October 17, 2026. The groups are lead and copper, arsenic, nitrate, total coliform and RTCR, disinfection byproducts, radionuclides, PFAS, and other health-based rules. The row names each group that had a violation, largest first. A county with none says so. A blank group is left off. The count of community systems is a separate number. People served is the total those systems with a violation reported. A system that sells water and a system that buys it can both carry population, and a system listed in two counties is counted in each. Adding those figures does not equal the number of county residents. The table does not call systems without a violation safe. (EPA SDWIS)
Well samples
A well-results table uses the state's own tested wells, one row per measure, with the limit the state names. In Michigan the rows are PFAS site investigations. The table is those tested wells or sites. It is not a prediction for a well the state did not test. The rows were checked October 9, 2026.
EPA radon zone
The water facts table names the county's EPA radon zone, links the EPA map of radon zones, and says how many of the state's counties share that zone. Zone 3 does not mean safe. EPA says test every home.
The county PFAS aggregation
Some county pages, and the guides for American Samoa, Guam, the Northern Mariana Islands, and the U.S. Virgin Islands, show a PFAS monitoring section from a county aggregation supplied on October 22, 2026. EPA published the underlying UCMR 5 occurrence data. EPA did not publish this county table. The final dataset was released in August 2026. EPA describes sample collection as January 2023 through December 2025, for 29 PFAS. Lithium is not in these figures. (EPA)
The count is public water systems associated with that place in the aggregation. It is not the share of homes. A system that serves two counties can appear in both, so the rows must not be added into a state, territory, or national total. A detection means a result at or above the study reporting level somewhere in that system's monitoring. It does not show that the sample was taken in the county, or that every household received that concentration. Private wells were not monitored.
Some recorded dates fall in 2026. Those dates are printed as recorded. This site has not confirmed that a 2026 date is the day a sample was collected. The file also does not say which rows used a ZIP-code fallback instead of a service-area match.
Where a page gives a PFOA or PFOS count at or above 0.004 µg/L, that is 4 ng/L. In drinking water, 4 ng/L is the same figure as the 4.0 ppt level EPA set for each of those two chemicals on April 10, 2024. The count is analytical results, not a violation. EPA's July 2026 data summary says UCMR 5 results do not indicate compliance or noncompliance. A PFBA, PFBS, PFPeA, PFHxS, or 6:2 FTS maximum is not that PFOA or PFOS count. (EPA data summary, July 2026)
A county with no row is a gap in this file. It is not proof that no system was sampled. When the older PFAS line in the water facts table disagrees with this aggregation, the table leaves that older line out so the page shows one system count. This page does not give a national percentage of counties sampled. The supplied file and the 2026 county list do not share one confirmed denominator.
Five rows cover part of American Samoa, Guam, the Northern Mariana Islands, and the U.S. Virgin Islands. Eastern District does not stand for all of American Samoa. Saipan does not stand for all of the Northern Mariana Islands. St. John has no row. Puerto Rico stays in the file and is not part of those four guides. Connecticut rows use the eight historic county codes, not the 2026 planning regions. One Alaska row still uses Valdez-Cordova Census Area, FIPS 02261, which the 2026 list no longer includes. That row is not copied onto Chugach or Copper River.
If a card looks wrong
Found a mistake? Tell us and we recheck the official page. Privacy, the terms, and the accessibility statement are on their own pages.
Each program's official page was opened between October 1, 2026 and November 16, 2026.